Incorrect license expression normalization for AND / OR combinations

[JustinWonjaePark]

Description

When ScanCode detects files with combined license expressions involving Apache-2.0 and MIT, the resulting license expression appears to be incorrectly normalized or simplified.

For a license statement such as

licensed under Apache-2.0 AND MIT

ScanCode reports only apache-2.0, omitting MIT entirely.

For a license statement such as

licensed under Apache-2.0 OR MIT

ScanCode reports the expression as: apache-2.0 AND (apache-2.0 OR mit)

How To Reproduce

scan files with statement as follows

licensed under Apache-2.0 AND MIT

and

licensed under Apache-2.0 OR MIT

scancode --license --json-pp scancode.json .

System configuration

OS: macOS 15.6.1 (x86_64)
ScanCode Toolkit version: 32.4.1
Installation method: pip

[Kaushik-Kumar-CEG]

I would like to take on this issue and resolve the bug!

[Dream-World-Coder]

Thanks for the clear report @JustinWonjaePark . I can reproduce this on my end. It seems like the composite license rule for Apache-2.0 AND MIT is missing or incorrectly defined in the rules database. I'll investigate the specific rule file and submit a PR to fix it.

[Kaushik-Kumar-CEG]

I have linked a PR to resolve this bug
#4691